Computer Vision Systems for “Context-Aware” Active Vehicle Safety and Driver Assistance - Digital Signal Processing for In-Vehicle Systems and Safety - page 204

Keywords Computer vision ￿ Context aware ￿ Lane tracking

15.1 Introduction

In their brief report, Fletcher et al. [ 1 ] provide an overall summary of promising
computer vision systems applied in the vehicle. They determine areas where vision
systems could be useful such as driver fatigue or inattention detection, pedestrian
spotting, blind-spot checking, lane keeping, traffic sign recognition, and human
factors aids. These applications are built based on several computer vision systems
which are surveyed and/or presented here in this chapter as well. Building on what
is already achieved in this area, we provide a systems engineering survey of
computer vision systems for in-vehicle applications together with our previous
and current findings. This study also presents a system utility analysis that ties all
systems in a mechatronics integration approach, reducing complexity and cost of
the final in-vehicle computer vision system, while maximizing the utility factor of
the resultant design. In Sect. 15.2 , applications are grouped into two main areas:
driver status monitoring (inside the vehicle) and vehicle peripheral monitoring
(outside the vehicle). These systems can be thought of as the eyes of the cyber
copilot in the vehicle which (or who) is aware of driver's condition as well as the
environment and the current situation (i.e., situation/context awareness). Next, in
Sect. 15.3 , all systems are analyzed from the perspective of utility in their projected
impact on reducing the number of accidents or fatality rates. After determining
the utility factors of systems, an example of mechatronics system integration
for in-vehicle systems is presented. Finally, conclusions are drawn in Sect. 15.4
pointing to future research directions in this area.

15.2 Computer Vision Systems for In-vehicle Applications

In this section, we briefly survey different CV systems, reporting our progress in
some areas with focus on the UTDrive research team. CV systems are seen as
crucial components of future DAS and AVS systems; however, there is still a need
for further development to achieve robust operation on board. Before providing
details on each system, a list of requirements from onboard CV systems are
presented here to emphasize the challenges in this area, some of which require
hardware solutions and development of novel systems:

￿ Robust against illumination change
￿ Reliable in vibration and high accelerations
￿ Durable to low/high temperatures and weather conditions (especially cabling
and mounting parts)
￿ Nonintrusive to the driver
￿ Compact/mobile
￿ Minimal power and computing source use